garden
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted project documentation and code during its scan phase to identify maintenance issues. * Ingestion points: Documentation files and source code are surveyed in Step 2 (SKILL.md). * Boundary markers: The instructions do not define strict boundary markers or "ignore instructions" delimiters when processing the scanned content (SKILL.md). * Capability inventory: The skill uses shell commands (git worktree, git commit, cd) and has the ability to write to the filesystem and open pull requests (SKILL.md). * Sanitization: No explicit sanitization or validation of the findings is performed before the agent implements the "fix" (SKILL.md).
- [COMMAND_EXECUTION]: The skill uses shell commands to manage git worktrees and navigate directories. Evidence:
REPO="$(git rev-parse --show-toplevel)",git worktree add "$WORKTREE" -b garden/<slug>, andgit worktree remove <path>(SKILL.md).
Audit Metadata